AlphaSignal reports that DeepSeek V4 Flash passed all 54 attempts in its private debugging test, comprising 18 tasks repeated three times. Its reported bill was $0.21, compared with $20.99 for Fable 5 at the same pass count. About 83% of V4 Flash's tokenThe chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.Full definition → usage received cache-read pricing, a major condition behind the cost difference. Benchmark account
The comparison measures bounded bugs graded against hidden tests and a regression suite. It supports a large reported cost advantage for those runs, while equal pass counts on that suite leave broader engineering capability unresolved.
Terms in this piece · Glossary
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
benchmark — A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
postPrompt cache TTL is the hidden line item in long coding-agent sessions
articleFull breakdown: Claude-only apps often wait until the host speaks the new revisi
articleDo not ask whether the agent follows the rule. Ask what stops it when it does no
articleFull Breakdown: When can a change ship unread? When something cheap and hard to 
